This paper presents a non-technical account of the developments in tree-based methods for the analysis of survival data with censoring. This review describes the initial developments, which mainly extended the existing basic tree methodologies to censored data as well as to more recent work. We also cover more complex models, more specialized methods, and more specific problems such as multivariate data, the use of time-varying covariates, discrete-scale survival data, and ensemble methods applied to survival trees. A data example is used to illustrate some methods that are implemented in R.

Tree‐based methods are frequently used in studies with censored survival time. Their structure and ease of interpretability make them useful to identify prognostic factors and to predict conditional survival probabilities given an individual's covariates. The existing methods are tailor‐made to deal with a survival time variable that is measured continuously. However, survival variables measured on a discrete scale are often encountered in practice. The authors propose a new tree construction method specifically adapted to such discrete‐time survival variables. The splitting procedure can be seen as an extension, to the case of right‐censored data, of the entropy criterion for a categorical outcome. The selection of the final tree is made through a pruning algorithm combined with a bootstrap correction. The authors also present a simple way of potentially improving the predictive performance of a single tree through bagging. A simulation study shows that single trees and bagged‐trees perform well compared to a parametric model. A real data example investigating the usefulness of personality dimensions in predicting early onset of cigarette smoking is presented. The Canadian Journal of Statistics 37: 17‐32; 2009 © 2009 Statistical Society of Canada

The COVID-19 outbreak has struck Lebanon in its worst period of instability, not only impacting physical health, but also increasing psychological distress. Using an online survey enhanced by response time measurement, this study describes the overall patterns in mental well-being outcomes and examines their association with sociodemographic characteristics during the COVID-19 pandemic. Furthermore, it identifies significant predictors for COVID-19 good practices. A total of 988 Lebanese were surveyed, with participants providing written online consent prior to filling the survey. Regression-based models were estimated. Findings show that individuals with higher education levels exhibit lower health concerns. People with children face higher health worries than those without. Men are more worried than women about their health and they are less satisfied with their lives during the pandemic. Descriptive statistics show that most Lebanese are very satisfied with their families (93.1%), but they are highly dissatisfied with their country (63%). Young adults and individuals who live alone exhibit significantly higher social well-being concerns. Age and having children were strong predictors for good COVID-19 practices. The odds of having good practices for older adults are 3.13 times higher than that of youth, while the odds for those with children are 3.18 times higher than those without. The findings of this study could pave the way for a well-coordinated national strategy and increased collaboration with public health professionals to mitigate the pandemic’s adverse effects on mental health in the long-term.

The aim of this paper is to propose a new survival tree method for discrete-time survival data with time-varying covariates. This method can accommodate simultaneously time-varying covariates and time-varying effects. The method is then used for bankruptcy analysis of US firms that conducted an Initial Public Offerings between 1990 and 1999 using accounting and financial ratios.

Purpose-This paper aims to assess the application of seven statistical and data mining techniques to second-stage data envelopment analysis (DEA) for bank performance. Design/methodology/approach-Different statistical and data mining techniques are used to secondstage DEA for bank performance as a part of an attempt to produce a powerful model for bank performance with effective predictive ability. The projected data mining tools are classification and regression trees (CART), conditional inference trees (CIT), random forest based on CART and CIT, bagging, artificial neural networks and their statistical counterpart, logistic regression. Findings-The results showed that random forests and bagging outperform other methods in terms of predictive power. Originality/value-This is the first study to assess the impact of environmental factors on banking performance in Middle East and North Africa countries.